The establishment of associations between structure and function among various areas of the brain is an important step in the identification of neurologic mechanisms in both normal and disease states. The pursuit of this goal requires the geometrical co-registration of digital images in two types of major applications: (1) data fusion, in the case brain images from the same subject were acquired by different modalities; (2) data comparison, for the detection of significant differences between different subject groups in images acquired by the same modality. The first application requires identification and delineation ("segmentation") of distinct areas and landmarks in the brain. A procedure based on dynamic clustering and region growing algorithms has been developed that segments T1-weighted MR images into regions of CSF, gray and white brain matter. This technique is currently being applied to the automatic/ semi-automatic identification of the following specific brain structures: cerebellum - good automated detection with operator allowed editing caudate - limited success at automated detection This technique in particular as been found to be of particular value in clinical research projects which use morphometric measures as an important measurement tool along with other physiological and sociological measures. Associated with the above developments, a technique based on a previously investigated algorithm has been further developed for automatic detection of axis of symmetry (- intra-hemispheric fissure) in PET and MR images. This algorithm is based on the correlation of similar but possible phased signals in the Fourier domain between values of corresponding spatial points. Using continuity conditions this has been extended to 3-D. This technique allows us to first identify the fissure in brain images and then using previously developed algorithms to orient the brain images of all subjects in the same vertical manner. This is helpful in obtaining consistent volumetric measures of brain structures which have traditionally been viewed as a 2-D cross- section. In addition, brain image visual presentation for both clinical review, editing purposes and landmark orientation for multimodality registration is also made more consistent. For the second (within-modality) application, the gray-level information itself can be employed for image registration without the need for segmentation. A multiscale registration procedure has been implemented that determines parameters of a general 3-D affine transformation (translation, rotation about an arbitrary center, anisotropic scaling and skewing) between volumes to be registered that minimizes the average squared gray-level difference between corresponding voxels. Successful registration of PET images achieving homogeneousregistration variance across the entire brain section has been achieved for both within and between subject analyses. This technique is now being used routinely in many of the clinical investigation. Furthermore this method is also being used for the co-registration of long time series (eg. 199 time points) of volumes acquired in functional MRI or fMRI studies.